In a search of new bone proteins it was discovered more than 10 years ago that YKL-40 was secreted in vitro in large amount by a human osteosarcoma cell line MG63 (Johansen et al. 1992). Two years later Morrison et al. (1994) reported that YKL-40 mRNA was expressed by murine mammary tumors initiated by neu/ras onco-genes but not by c-myc or int-2 oncoonco-genes. Although MG63 cells originate from an osteosarcoma, these cells also have chondrocyte characteristics, since their metastases mainly consist of proliferating nodules of hypercellular cartilage (Heremans et al. 1978). Further-more, unstimulated MG63 cells synthesize larger amount of type III collagen than of type I collagen and secrete low to undetectable levels of alkaline phosphatase and osteocalcin (Franceschi et al.
1988). Today it is known that many different types of human solid cancer express YKL-40. A search of the YKL-40 sequence against the dbest database at the National Center for Biotechnology Informa-tion showed that several types of solid cancer (breast-, colon-, lung-, kidney-, pancreas-, ovarian-, prostate-, and uterine carcinoma, osteosarcoma, oligodendroglioma, glioblastoma and germ cell tu-mors) overexpressed YKL-40. Microarray gene analyses have identi-fied the YKL-40 gene to be one of the most highly over-expressed genes in high-grade malignant gliomas (Lal et al. 1999; Markert et al. 2001; Tanwar et al. 2002), in papillary thyroid carcinoma (Huang et al. 2001), and in extracellular myxoid chondrosarcoma (Sjögren et al. 2003). YKL-40 is not expressed by myxoid liposarcomas (Sjö-gren et al. 2003). Shostak et al. (2003) used public databases of the Cancer Genome Anatomy Project and found enhanced expression of the YKL-40 gene in glioblastoma multiforme and occasionally in anaplastic astrocytomas compared to normal brain. The upregula-tion of YKL-40 in glioblastoma multiforma is also confirmed on the protein level by Western blotting where 65% of the investigated glio-blastoma multiforma samples had stronger YKL-40 protein expres-sion than low-grade gliomas (Tanwar et al. 2002). YKL-40 is also se-creted in vitro by human glioblastoma cells (Junker et al. 2005b) and by the monocyte-like human histiocytic lymphoma cell line U937 (Verhoeckx et al. 2004).
The biological function of YKL-40 in cancer diseases is unknown.
It has been hypothesized that YKL-40 is a growth factor of cancer cells or protects them from undergoing apoptosis. YKL-40 is also called the “breast regression protein (Brp-39)” (Morrison et al.
1994) because it is induced in mice mammary epithelial cells a few days after weaning. Mammary involution involves programmed cell death, and it has been hypothesized that YKL-40 utilizes a chitin oligosaccharide binding ability while participating in various signal transduction pathways leading to apoptosis of regressing cells, and that YKL-40 is a protective signaling factor determining which cells are to survive the drastic tissue remodeling that occurs during in-volution (Mohanty et al. 2003). Cancer cells that express YKL-40 may have a different phenotype compared to cancer cells without YKL-40 expression, and the protein may reflect differences in the biology of various cancer cells. Neoplasms are biologically hetero-geneous and contain subpopulations of cancer cells with different angiogenic, invasive and metastatic properties. Metastases can have a clonal origin, and different metastases can originate from the pro-liferation of different single cells. The process of metastases is select-ive for cells that succeed in promoting angiogenesis, invasion, em-bolisation, survival in the circulation, arrest in distant capillary beds and extravasation into and multiplication within the organ par-enchyma. It need to be determined if YKL-40 has a role in one of these processes.
Cancer progression depends on the interplay between the cancer cells and their micro-environment, particularly the surrounding ECM, and the balance between synthesis and degradation of ECM components is a key modulator of cancer growth and metastasis (Boudreau et al. 1998). The stroma around the periphery of solid cancers have several similarities with granulation tissue such as that found in wound-healing or inflammation (Dvorak 1986; Gregoire et
al. 1995), and tumors are called “wounds that never heal” (Balkwill et al. 2001). Recent studies have shown that tumor-associated macrophages and leukocytes play important roles in tumor growth and metastasis, since these cells produce growth and angiogenic fac-tors, chemokines, chemotactic facfac-tors, MMPs and other ECM de-grading enzymes (Sunderkötter et al. 1994; Lin et al. 2001, 2004ab;
Bingle et al. 2002; Pollard 2004). Ongoing immunohistochemical analysis of YKL-40 expression in biopsies from breast cancer (Anne Roslind; manuscript in preparation) and colorectal cancer (personal observation) show that cancer cells in some biopsies have YKL-40 protein expression. Furthermore macrophages and neutrophils in the stroma surrounding the breast and colorectal cancer cells have YKL-40 protein expression. In situ hybridization of YKL-40 mRNA expression in biopsies from small cell lung cancer shows no YKL-40 mRNA expression in the cancer cells but strong expression in peri-tumoral macrophages (Junker et al. 2005a). It is unknown if these macrophages are CD14+,CD16+, a phenotype that express YKL-40 in RA patients (Baeten et al. 2000) and is increased in number in pa-tients with solid cancers (Saleh et al. 1995).
YKL-40 purified from the MG63 osteosarcoma cell line has growth factor activity for fibroblast cell lines (Recklies et al. 2002).
One could speculate that YKL-40 secreted by cancer cells and tu-mor-associated macrophages and neutrophils has a role in prolifer-ation, activation and differentiation of the fibroblasts/myofibro-blasts surrounding the tumor, and thereby influence development of the prominent desmoplastic fibroblast stroma seen in both primary cancer and metastatic sites. The phenomenon, termed stromal reac-tion, includes activation of fibroblast or myofibroblastic transfor-mation, enhanced secretion of matrix proteins and MMPs, and neo-vascularization all of which promote proliferation, differentiation, invasion or regression of cancer cells and destruction of the stroma (Dvorak 1986; Basset et al. 1990; Gregoire et al. 1995; Rønnov-Jes-sen et al. 1996; Bissell et al. 2001; Kenny et al. 2003).
YKL-40 also stimulates migration of endothelial cells at a level comparable to that achieved by bFGF (Malinda et al. 1999) and modulates vascular endothelial cell morphology by promoting the formation of branching tubules. YKL-40 may therefore be a positive regulator of angiogenesis surrounding the tumor and could play a role in the growth of primary and metastatic tumors. Junker et al.
(2005b) found upregulated YKL-40 expression in a human glioblas-toma cell line by genotoxic and micro-environmental stress (i.e.
exposure to hypoxia, ionizing radiation, etoposide, ceramide, p53 inhibition, antioxidant treatment, confluence, and serum deple-tion). The response in YKL-40 expression was late, 24-72 hours after stimuli, indicating that YKL-40 is a secondary response downstream of other mechanisms.
One can therefore speculate that YKL-40 is involved in prolifer-ation of cancer cells, the surrounding tissue remodeling processes and angiogenesis, and that serum concentrations of YKL-40 may be a novel “Tumor marker”. The term ”Tumor marker” embraces a spectrum of molecules of widely divergent characteristics (e.g. cyto-genic markers, oncogenes and abnormally expressed proteins with various biological functions), but sharing an association with malig-nancy that facilitates their application in the clinical detection (diag-nosis, screening) and management (monitoring, prognosis) of can-cer patients. Tumor markers are biological compounds, produced either by tumor cells or by the host in response to a developing tumor and are usually determined in serum. A large number of pro-teins have been suggested as potential circulating “Tumor markers”
(Sturgeon 2002): e.g. 1) serum carcinoembryonic antigen (CEA) (Hayes et al. 1996; Mitchell 1998; McLeod et al. 1999; Compton et al. 2000; Thomas et al. 2001; Duffy et al. 2003) and plasma TIMP-1 (Holten-Andersen et al. 2000, 2002; Duffy et al. 2003) in colorectal cancer; 2) serum CA-125 (Bast et al. 1998, 2003) and tetranectin (Høgdall et al. 2000a) in ovarian cancer; 3) serum prostate specific antigen (PSA) in prostate cancer (Catalona 1994; Partin et al. 1997;
Canto et al. 2004; Hittelman et al. 2004; Khan et al. 2004); 4) serum
alpha-fetoprotein (AFP) and human chorionic gonadotropin (hCG) in testis cancer (Bosl et al. 1997); and 5) serum breast cancer as-sociated antigen 549 and 15.3 (Söletormos 2001), plasma soluble urokinase plasminogen activator receptor (Riisbro et al. 2002) and serum human epidermal growth factor receptor 2 (HER-2) in breast cancer (Ross et al. 1998; Carney WP 2003). None of these bio-markers are specific for cancer and are not yet used routinely in screening for cancer. Serum CEA, CA-125, PSA, AFP and hCG are applied routinely in patients suspected of having cancer and in monitoring of cancer patients.
IS SERUM YKL-40 A NEW BIOMARKER IN CANCER PATIENTS?
Acceptance of novel tumor markers in clinical settings requires thor-ough validation before being implemented into routine clinical use.
Werner et al. (1993) have suggested that “Tumor markers” are classi-fied according to six different clinical criteria such as biochemical characteristics, organ specificity or clinical usefulness in order to as-sess the value of tumor markers in clinical practice:
1. “The marker is produced exclusively by specific tumor cells (tumor specific)”?
40 is not specific for cancer or a certain type of tumors. YKL-40 is produced by non-malignant cells (as described in Chapter 1 and 4) and by cancer cells of widely different types of solid cancer.
2. “The marker is absent in healthy or benign disease (high specificity)”?
Serum concentrations of YKL-40 do not have high specificity for cancer. YKL-40 is detected in serum from healthy subjects and ele-vated serum YKL-40 (compared to healthy subjects) are found in patients with non-malignant diseases such as severe bacterial infec-tions, active RA, GCA, IBD, lung sarcoidosis, and liver fibrosis (as described in Chapter 4). Most patients with these diseases will have some clinical symptoms of their disease.
3. “The marker is present frequently in the targeted malignancy (high sensitivity)”?
In 1995 Johansen et al. (II) reported that some patients with meta-static breast cancer had increased serum YKL-40 compared to healthy subjects, and that the highest serum YKL-40 were found in patients with short survival (Figure 4). This study suggested that serum YKL-40 might be useful as a prognostic marker in breast
cancer patients. Recent studies have found elevated serum YKL-40 in a subgroup of patients with seven different types of localized or metastatic solid cancer compared to healthy subjects (Table 3) (Cin-tin et al. 1999, 2002; Tanwar et al. 2002; Dehn et al. 2003; Geertsen et al. 2003; Høgdall et al. 2003; Jensen et al. 2003; Johansen et al.
2003a, 2004; Dupont et al. 2004; Brasso et al. 2006). It needs to be determined if YKL-40 is elevated in serum of patients with hemato-logical malignancies.
Preoperative serum YKL-40 levels were elevated in 19% of pa-tients with primary breast cancer and the papa-tients with metastases to axillary lymph nodes had higher serum YKL-40 compared to lymph node negative patients (Johansen et al. 2003a). In patients with first recurrence of breast cancer serum YKL-40 was elevated in 31-41%
of the patients, and high serum YKL-40 was associated with meta-static sites and large tumor load: 9-20% of patients with recurrence to lymph nodes or skin only had elevated serum YKL-40, 24-35%
with bone metastases, and 57-61% with visceral metastases (Jo-hansen et al. 1995; Jensen et al. 2003). Highest serum YKL-40 were found in patients with more than two different metastatic sites (Jensen et al. 2003). Preoperative serum levels of YKL-40 from patients with colorectal cancer was elevated in 26% and there was an association between serum YKL-40 and Dukes’ stage: 16% of the patients with Dukes’ A (tumor confined within the bowel wall, no lymph-node metastases), 26% with Dukes’ B (tumor extending through the bowel wall, no lymph-node metastases), 19% with Dukes’ C (regional lymph-node metastases), and 39% with Dukes’
D disseminated disease) had elevated preoperative serum YKL-40 (Cintin et al. 1999). Serum YKL-40 decreased significantly after curative operation for colorectal cancer in patients with high pre-operative serum YKL-40 (Cintin et al. 2002), indicating that serum YKL-40 reflect tumor burden. Preoperative serum YKL-40 was ele-vated in 65% of stage I and II ovarian cancer patients (Dupont et al.
2004), in 74-91% of patients with ovarian cancer stage III (tumor growth involving one or both ovaries with wide-spread intraperito-neal metastases) and IV (disseminated disease) (Høgdall et al. 2003;
Dupont et al. 2004) and in 55% of ovarian cancer patients at time of first recurrence (Dehn et al. 2003). In patients with small cell lung cancer 22% with local disease and 40% with extended disease had elevated serum YKL-40 (Johansen et al. 2004). 43% of patients with metastatic prostate cancer (Brasso et al. 2006) and 83% of patients with metastatic renal cell cancer (Geertsen et al. 2003) had elevated serum YKL-40. In patients with glioblastoma serum YKL-40 was re-lated to tumor grade and burden: 72% of patients with glioblastoma multiforme and 57% with lower grade gliomas had high serum YKL-40 (Tanwar et al. 2002).
These studies demonstrate that serum YKL-40 does not have a high sensitivity for solid carcinoma, suggesting that not all tumors express YKL-40 or it is secreted at a low level. This could be evalu-ated in immunohistochemical and in situ hybridization studies of biopsies from different types of cancer.
4. “The marker is detectable in early stage subclinical disease (useful for screening)”?
Determination of serum YKL-40 concentrations cannot be used as a single screening test for cancer. At time of first cancer diagnosis 16-74% of the patients had elevated serum YKL-40, and only 16-26% of patients with primary localized cancer had elevated serum YKL-40.
However, in patients with locally advanced or metastatic cancer at the time of diagnosis serum YKL-40 levels were elevated in 39-83%.
A high serum YKL-40 in a subject without any known disease may therefore indicate non-symptomatic cancer. Serum concentrations of YKL-40 were independent of serum CEA in colorectal cancer pa-tients (Cintin et al. 1999, 2002), of serum CA-125 in ovarian cancer patients (Dehn et al. 2003; Høgdall et al. 2003), of serum HER-2 in metastatic breast cancer patients (Jensen et al. 2003), of serum LDH in patients with small cell lung cancer (Johansen et al. 2004) and of serum PSA in patients with metastatic prostate cancer (Brasso et al.
Figure 4. Individual serum YKL-40 concentrations in patients with meta-static breast cancer in relation to months of survival after the serum sample was obtained (Johansen et al. 1995 II). The serum YKL-40 levels were de-termined by RIA (Johansen et al. 1993 I) but the data were corrected to ELISA values (YKL-40 ELISA = YKL-40 RIA X 0.479). The upper 95thpercent limit of serum YKL-40 in 245 healthy adults is 124 µg/l.
Survival (months)
0 12 24 36 48 60 72
0 100 124 200 300 400 500 600
Serum YKL-40 (µg/l)
2006). The studies indicate that serum YKL-40 reflects other aspects of tumor growth and metastasis than these tumor markers. It may be of value to include serum YKL-40 as a biomarker for screening of cancer together with a panel of other tumor markers and imaging techniques, since an elevated serum YKL-40 level seems to reflect metastatic disease and secretion from a subset of tumors with a more aggressive phenotype and with a poor prognosis (described below).
5. “The marker’s concentration reflects prognosis for an individual patient (prognosticator)”?
Eight studies have demonstrated that elevated serum YKL-40 in pa-tients with breast-, colorectal-, ovarian-, kidney-, small cell lung-, and prostate carcinomas was an independent prognostic parameter of short recurrence free interval and short overall survival with hazard ratios between 1.3 and 4.1 (Table 4). This observation was found in patients with local or metastatic cancer, and at the time of
first cancer diagnosis or at the time of relapse (Johansen et al. 1995, 2003a, 2004; Cintin et al. 1999, 2002; Dehn et al. 2003; Geertsen et al. 2003; Høgdall et al. 2003; Jensen et al. 2003; Brasso et al. 2006).
High preoperative serum YKL-40 level in patients with primary breast cancer was an independent prognostic parameter of short re-currence free interval and short overall survival when axillary lymph node- and estrogen receptor status, age, tumor size and histology, menopausal status and serum YKL-40 were included in the multi-variate Cox analysis (Johansen et al. 2003a). There are no longitu-dinal studies of the changes in serum YKL-40 levels in breast cancer patients after operation and adjuvant chemo-, antiestrogen- or radio-therapy. However, an elevated serum YKL-40 in breast cancer patients at time of first recurrence predicted shorter time to progres-sion and shorter overall survival (Jensen et al. 2003). Multivariate Cox analysis (including estrogen receptor- and axillary lymph node status at primary diagnosis, liver metastases, more than two meta-static sites, symptomatic disease at recurrence and serum HER-2
Relative 95%
hazard confidence
Diagnosis ratio interval p value Reference
Primary breast cancer . . . . 1.8 1.0-3.1 0.04 Johansen et al. 2003a Metastatic breast cancer . . . . 2.6 1.6-4.1 0.0002 Jensen et al. 2003 Colorectal cancer . . . . 1.4 1.1-1.8 0.007 Cintin et al. 1999 Ovarian cancer stage III . . . . 4.0 1.5-10.3 0.005 Høgdall et al. 2003 Recurrent ovarian cancer . . . . 2.3 1.3-4.1 0.006 Dehn et al. 2003 Small cell lung cancer . . . . 1.9 1.1-3.4 0.02 Johansen et al. 2004 Metastatic prostate cancer . . . . 1.3 1.0-1.7 0.02 Brasso et al. 2006 Metastatic renal cell cancer . . . . 4.1 1.9-8.8 0.001 Geertsen et al. 2003 These cancer patients were scored as having elevated serum YKL-40 if it was higher than the upper 95thper cent confidence limit of serum YKL-40 in healthy subjects adjusted for age (se also footnotes to Table 3).
Table 4.Serum level of YKL-40 is an independent prognostic variable of overall survival in cancer patients. These results are from multivariate Cox regression analysis using routinely used prognostic variables.
High YKL-40
Diagnosis N Serum YKL-40 (%)# Reference
Primary breast cancer¤ 271 57c (22-688) 19 Johansen et al. 2003a
Metastatic breast cancer§, relapse . . . . 54 80c (20-560) 41 Johansen et al. 1995 soft tissue . . . . 10 59 (29-433) 20
bone . . . . 25 75c (21-560) 35 viscera . . . . 19 157c (20-468) 61
Metastatic breast cancer, 1. relapse . . . 100 65c (20-430) 31 Jensen et al. 2003 nodes and skin only . . . . 36 51 (20-267) 9
bone . . . . 28 61c (24-310) 24 viscera . . . . 36 110c (21-430) 57
Colorectal cancer ¤ § . . . 603 86c (27-1298) 26 Cintin et al. 1999 Dukes A . . . . 58 73b (27-295) 16
Dukes B . . . 223 86c (27-604) 26 Dukes C . . . 175 77c (27-582) 19 Dukes D . . . 147 119c (27-1298) 39
Glioblastoma multiforme . . . . 45 130c (38-654) 72 Tanwar et al. 2002 Lower grade gliomas . . . . 20 101c (50-225) 57
Ovarian cancer, all stage s¤ . . . . 50 94c (17-517) 72 Dupont et al. 2004 Ovarian cancer, stage III¤ . . . . 47 168c (32-1808) 74 Høgdall et al. 2003 Ovarian cancer, relapse . . . . 73 94c (20-1970) 55 Dehn et al. 2003 Small cell lung cancer§ . . . 131 82c (23-1188) 32 Johansen al. 2004
local disease . . . . 59 71a (23-417) 22 extensive disease . . . . 72 101c (27-1188) 40
Metastatic prostate cancer 153 112c (20-2080) 43 Brasso et al. 2006
Metastatic renal cell cancer 58 235c (45-1896) 83 Geertsen et al. 2003
Values are median (range).
a: p < 0.02, b: p < 0.01 and c: p < 0.001, compared with controls (Mann-Whitney test).
¤) Preoperative.
#) The percentage (%) of patients with elevated serum YKL-40 compared to the age-adjusted serum YKL-40 level in healthy subjects. For all the danish studies the normal reference region was calculated on the log transformed serum or plasma YKL-40 levels obtained from healthy subjects (aged 18-79 years; N = 260 for RIA values and N = 245 for ELISA val-ues) (Johansen et al. 1996a III). The upper 95thper cent confidence limit was chosen for the limit and adjusted for age (Royston 1991).
§) RIA analysis (Johansen et al. 1993 I) but data corrected to ELISA values (YKL-40 ELISA = YKL-40 RIA X 0.479). All the other studies used the ELISA method (Harvey et al. 1998).
Table 3. Serum levels of YKL-40 (µg/l) in patients with localized or advanced cancer and the percentage of patients with elevated serum YKL-40.
and YKL-40 levels) showed that high serum levels of YKL-40 and HER-2 were independent prognostic variables of short time to dis-ease progression and death (Jensen et al. 2003). Figure 5 illustrates survival curves in patients with metastatic breast cancer according to elevated or normal serum concentrations of YKL-40 and HER-2 at time of first relapse (Jensen et al. 2003). Patients with both high serum YKL-40 and HER-2 level had the poorest median survival of only 9 months contrasting 32 months for patients with normal serum YKL-40 and HER-2.
High preoperative serum concentration of YKL-40 in patients with colorectal cancer was also an independent prognostic para-meter of short recurrence free interval and short overall survival (the multivariate Cox analysis included Dukes’ stage, age, gender, serum CEA and YKL-40) (Cintin et al. 1999). In stage III ovarian cancer patients a high preoperative serum YKL-40 was an independ-ent prognostic parameter of short survival (the multivariate Cox analysis included serum YKL-40 and CA-125, optimal vs. subopti-mal results from primary surgery, age, and histological type of tu-mor) (Høgdall et al. 2003) and a similar result was found in patients with recurrence of ovarian cancer (the multivariate Cox analysis in-cluded serum YKL-40 and CA-125, age, localization of tumor and its size, performance status, primary and second-line treatment) (Dehn et al. 2003). An elevated serum YKL-40 was also an inde-pendent prognostic variable of short survival in patients with meta-static prostate cancer (the multivariate Cox analysis included age, performance status, tumor grade, serum PSA, total and bone alka-line phosphatase, PINP, crosslaps and YKL-40) (Brasso et al. 2006), and in patients with metastatic renal cell carcinoma (the multivari-ate Cox analysis included serum YKL-40, performance status, number of organ disease sites, organ site involvement, prior ne-phrectomy, and time from diagnosis to metastases) (Geertsen et al.
2003). In patients with small cell lung cancer a high serum YKL-40
2003). In patients with small cell lung cancer a high serum YKL-40